[PATCH 5/5] Drivers: hv: vmbus: Implement the protocol for tearing down vmbus state

[Vitaly Kuznetsov]

Vitaly Kuznetsov <vkuznets at redhat.com> writes:

> "K. Y. Srinivasan" <kys at microsoft.com> writes:
>
>> Implement the protocol for tearing down the monitor state established with
>> the host.
>>
>> Signed-off-by: K. Y. Srinivasan <kys at microsoft.com>
>
> Unfortunatelly this patch leads to the following crash:
>
> pre-udev:/# modprobe hv_vmbus
> modprobe hv_vmbus
> [   14.832334] hv_vmbus: Hyper-V Host Build:9600-6.3-17-0.17039; Vmbus version:3.0
> [   14.834179] hv_vmbus: CPU offlining is not supported by hypervisor
> pre-udev:/# modprobe -r hv_vmbus
> modprobe -r hv_vmbus
> [   20.640893] BUG: unable to handle kernel NULL pointer dereference at           (null)
> [   20.641104] IP: [<          (null)>]           (null)
>
> I'll try to investigate.
>

The problem is that we call vmbus_disconnect() too late in
vmbus_exit(). We need to call it:
1) Before hv_cleanup() call as we free hypercall page there (that's what
you see in my crash)
2) Before hv_remove_vmbus_irq() as we're waiting for the hypervisor to
reply.

This simple patch fixes the issue:
diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c
index 7870a90..2b56260 100644
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -1106,6 +1106,7 @@ static void __exit vmbus_exit(void)
 
        vmbus_connection.conn_state = DISCONNECTED;
        hv_synic_clockevents_cleanup();
+       vmbus_disconnect();
        hv_remove_vmbus_irq();
        vmbus_free_channels();
        if (ms_hyperv.features & HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE) {
@@ -1118,7 +1119,6 @@ static void __exit vmbus_exit(void)
                smp_call_function_single(cpu, hv_synic_cleanup, NULL,
        1);
        acpi_bus_unregister_driver(&vmbus_acpi_driver);
        hv_cpu_hotplug_quirk(false);
-       vmbus_disconnect();
 }

I suggest we incorporate it into your PATCH 5/5.

-- 
  Vitaly